JPS63200100A - Method and device for automatically operating nuclear power plant - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to the automation of the operation of nuclear power plants, and is particularly applicable when it is necessary to change the operation plan in order to satisfy the operation restriction conditions of the nuclear reactor. The present invention relates to an automated method and an automated device for automatically determining a corrected road curve and setting a driving target.

[Conventional technology]

As a known technology whose main objective is to automate operation plan creation, output control, core monitoring, and operation plan correction during operation of a nuclear power plant, for example, Japanese Patent Application Laid-Open No. 57-1353
There are 97. In this known example, (a) creation of an operation plan that uses a core simulator to automatically search for a load pattern that is closer to the requested load pattern (equal to the basic load curve) from the central power dispatch center and that complies with the operation restriction conditions; (2) Captures plant data from the reactor and outputs an operation guide request command when it is determined that margins for operation restriction conditions (core flow rate, linear power density, etc.) exceed preset allowable values. According to the core condition monitoring device and (c) the operation planning device, the core condition after a short period of about one hour and margins for the operation restriction conditions are predicted and calculated, and if the operation is performed according to the operation plan, it will be determined that the operation restriction conditions will be violated. and a driving guide device that formulates a temporary driving method by converting the ratio between the predicted value and the driving limit value into output and subtracting it from the planned value.
A reactor load following control system is presented.

[Problem that the invention seeks to solve]

In the above-mentioned known example, only the case where there is no margin in the driving restriction conditions is taken up as a factor for modifying the road curve determined from the driving plan. However, even if there is a margin in the operating limit conditions, when the reactor output is changed significantly, for example at plant start-up, it is not always possible to adjust the control rod settings due to errors in the core simulator used to predict the reactor state. The target value of reactor power/core flow rate may not be achieved with the pattern. In other words, the simulators used always have errors, even if there are differences in degree, so it seems necessary to create and recalculate detailed rotor curves.

Additionally, in a known example, when it is determined that driving as planned would violate the driving restriction conditions, a temporary driving method is devised by converting the ratio between the predicted value and the driving restriction value into output and subtracting it from the planned value. However, as a result of local changes in the driving method, the subsequent driving plan also needs to be significantly revised, so there is a risk that the result will deviate significantly from the required basic load curve. To prevent this, if it is necessary to change the operating method, a modified load should be applied across the entire load curve so that the deviation from the required reactor output according to the basic load curve is small and the margin for the operating limit conditions is large. A method that creates a curve and determines a new driving method based on it is desirable.

The purpose of the present invention is to provide a function to address the two problems described above, namely (i) error correction of the core simulator used for prediction, and (n) loading when there is a risk of violation of the simulator's prediction error or operational restriction conditions. An object of the present invention is to provide a method and an automation device for automating the operation of a nuclear power plant, which enables flexible response to the actual operating conditions of a nuclear reactor by providing a function for optimally correcting a curve.

[Means for solving problems]

The above objective can be achieved by first comparing the actual measured values of the plant operating parameters and the predicted values by the core simulator every hour and creating an error prediction model through statistical processing, thereby creating a prediction error model. Next, if the prediction error is larger than a predetermined value or if it is predicted that the operation restriction conditions will be violated, the deviation from the requested basic load curve and the margin of the operation restriction conditions are respectively set in the output fluctuation operation section. By creating an optimal correction lobe curve using an evaluation function whose parameter is the value integrated over the entire driving plan, it is possible to appropriately change the driving method considering the entire driving plan.

[Effect]

The plant operating parameters obtained by predictive calculation using a core simulator are (A) [those obtained by a lumped constant model such as the average output of the child reactor and the average void ratio of the core, and (B)
There are things that depend on the distribution model of the reactor core, such as power distribution and thermal limit values. The former parameter (A) shows a locally approximately linear variation response to the output variation according to the load curve.

The latter parameter (B) also depends on changes in the control rod pattern, so it does not necessarily indicate a continuous fluctuating response that can be described by a linear model.However, there are also conditions under which the control rod pattern can be considered constant or similar. Local linearization can be achieved when changing the output by adjusting the core flow rate. Therefore, under the above uniform operating conditions, if we continuously calculate the difference between the actual measured value and the predicted value at that point determined in advance by a simulator, the time series will be expressed in an appropriate linear form or at most 2 It can be described by the following formula.

Error correction of the simulator can be realized by creating a statistical prediction model for future changes in such time series using techniques such as multivariate analysis.

On the other hand, as a function of operating parameters (core flow rate, control rod pattern, etc.) at each point of the modified load curve, the deviation from the required basic load curve and the margin of operational limit conditions can be defined. ``A rotation limit condition can be used simply as a constraint condition in optimal search to solve the optimization problem of a modified load curve, but in general, it is used as a parameter using the reciprocal of the difference between the normalized operation limit conditions and the limit value. The former nonlinear optimization problem with constraints can be solved by adopting, for example, the effective permissible direction method.The latter generalized optimization problem can be solved using the following method: This is a multi-objective programming problem in which a vector objective function is minimized under given constraints, but in order to comprehensively evaluate it, we use the absolute value of the vector as the evaluation function. It is possible to solve it.

〔Example〕

Embodiments of the nuclear power plant operation automation device of the present invention will be described below. FIG. 1 shows the eight functional configurations of the embodiment, which can be roughly divided into a plant 3 and a manipulated variable for controlling the control variables of the plant 3 to be equal to the target value. Subloop control device that outputs! !
4. A general control device 5 that outputs a target value to the subloop control device, a state prediction section 6 and a state monitoring section 7 that collectively form the function of outputting a modified lobe curve to the general control device.
An operation plan correction section consisting of a simulator error correction section 8 and a curve optimization correction section 9, and an operation request command from an operator are input to the integrated control device 5, and the operation state of the plant and calculation results of each partial function of the operation automation device are inputted. It consists of a man-machine interface 2 that is presented to the operator. Hereinafter, the operation of the operation plan modification section, which is the main technical feature of the present invention, will be described in detail.

The condition monitoring unit 7 takes in the plant data 11 and uses a core monitoring device using a core performance calculator, which is a technology already adopted in current plants, to quickly determine operating parameters such as core power distribution and thermal limit values. The calculation is performed and compared with the driving limit value from the driving plan creation knowledge database 10 to determine the degree of margin of the driving restriction condition. The state prediction unit 6 takes in the plant data 11 and uses a core simulator to obtain predicted operating parameter values at each point of the curve when the operating target for output fluctuation is set and the operation is performed according to the load curve. The simulator error correction unit 8 takes in the measured values of the operating parameters obtained by the condition monitoring unit 7 and the predicted values of the operating parameters corresponding to these measured values and obtained in advance by the condition prediction unit 6, and calculates these values. Calculate the correction value for the simulator error using the time series of differences.

The curve optimum modification unit 9 takes in the margin of the driving restriction conditions and the correction value of the simulator error, determines whether or not it is necessary to modify the load curve currently used, and determines that it is necessary to modify it. In this case, a modified road curve is created using the driving plan creation knowledge database 10.

The basic calculation procedure of the operation plan modification section described above is shown in FIG. 2. Below, detailed calculation contents in simulator error correction 8' and curve optimization correction 9', which process the calculation steps encircled by the outline, will be described.

FIG. 3 shows the calculation procedure in the simulator error correction section 8. The prediction error time series is calculated using the following formula.

δxD)=x(t)-x(t)...(
1) (t=to, ttt...shinn) Here, x(t)'' (xt(t), xz(t), xq
(t)) are actual measured values of operating parameters XI, X2.・・・
・A vector whose elements are
, -tn) is a time series of prediction errors. The basic assumption adopted for the statistical model representing the time series of prediction errors is that these time series contain a systematic variable constraint as represented by a linear form of some explanatory variable. If systematic fluctuation factors are not included, the magnitude of the fluctuation is equivalent to the variance of stochastic fluctuations, so it can be ignored as a prediction error. When a multivariate multiple regression model is adopted as the statistical model and the time series of prediction errors is taken as the objective variable, the following relational expression can be applied.

δx (t) = β ・u (t) + ε (1)
...(2) (t=to, ttt...tn) Here, β is a partial regression coefficient matrix CPXq).

u(t) is the explanatory variable vector (P dimension), ε (1)
is the residual vector (q dimension). The explanatory variables adopted in this case include the manipulated variables that are changed when operating the plant according to the load curve, such as the core flow rate,
The number of steps in the control rod withdrawal sequence, the Xe concentration, etc. may be determined. Therefore, the least squares estimated value β of the partial regression coefficient matrix Δβ is obtained using the method of multivariate multiple regression analysis, and using this, the current tn
The predicted result correction value x umbrella (tz) of the simulator at a future time tz is expressed by the following formula (tffi>tn). Here, equation (3) giving the predicted result correction value sets the independent variable to a future time t. ! However, this is the way to express it because of the expression format, and the actual independent variable is the explanatory variable vector U. This means that the model equation (2)
The time parameter tl in is merely to indicate the meaning of the numbering of individual data. Also 1
, From this, equation (3) is a relational equation that holds true for each scalar component of the predicted result correction value vector, and therefore, for a load curve different from that used to create model equation (3). is also applicable.

The method of correcting simulator errors described above is an example of a statistical analysis method that does not rely on a physical model, but there are other methods that feed back actual measured values to the physical model that has been widely used in the past. This method allows simultaneous use of a method based on tuning of simulator input data and a method based on tuning of simulator input data.

FIG. 4 shows a detailed calculation block diagram of the curve optimum correction section 9. The modified road curve assumption unit 21 takes in the margin of the current driving restriction condition from the condition monitoring unit 7 and the correction value δx-(t)=βu(shi) for the prediction error difference from the simulator error correction unit 8, respectively. First, it is determined whether there is a need to newly modify the load curve currently in use.

As a criterion for determination, if there is no margin for the driving restriction conditions or if the correction value of the prediction error of the simulator is larger than a certain reference value set, the load curve correction calculation is started.

Next, when starting correction calculations, the first assumption is that the load curve currently in use is used as the corrected load curve, but if no correction has been made at this point, data input from an external source will be used. A basic load curve 27 given as is used.

The input processing unit 22 takes in the assumed corrected road curve, and calculates the following for the main operating points characterizing the road curve.
Create input data for simulation to obtain predicted values of operating parameters (output distribution, thermal limit values, etc.) and process the data.

In the predictive simulation 23, input data for simulation is input to the core simulator 28, and predicted values of operating parameters at major operating points of the load curve are determined. At this time, information regarding correction of the simulator model and input data tuning from the simulator error correction unit 8 is reflected. Next, the result extraction unit 24 extracts the predicted values of the operating parameters from the simulation results, and corrects the predicted values of the simulator using the predicted result correction value (Equation (3)) given from the simulator error correction unit 8. The curve performance evaluation unit 25 takes in the predicted driving parameter values at each point of the road curve, and evaluates the validity of the road curve using an evaluation function as shown in the following equation, for example.

5ubj, to uεU= (uig(u)≦0), where f (u)= (f x(u), f z(u)
, ・= f h(u)) is the evaluation function vector,
The elements f1...fb are used to evaluate the appropriateness of the load curve, such as the deviation of various operating parameters from the target value, the margin of thermal limit value, the goodness of output distribution, the number of changes in control rod pattern, etc. Parameters are standardized and weighted. g(u)=(gt(u),...gr(u
)) is the constraint condition vector, and UεR' is the torque between the operation amount and operating conditions when driving according to the load curve, and can be taken as the same as the explanatory variable vector used in the simulator's prediction error model (Equation (2)). . The optimum measure search unit 26 uses the road curve evaluation function created by the curve performance evaluation unit 25 to search for an optimum direction for road curve correction. Whether the modified load curve is optimal can be determined by whether the evaluation function no longer decreases.

When the optimal point is reached, the optimally corrected load curve is outputted to the integrated control unit 5, as well as information on vectors of operating amounts and operating conditions for operating the plant in accordance with the optimally corrected load curve. Furthermore, if a new corrected road curve is obtained through the optimal direction search, the data is sent to the corrected road curve assumption section 21,
The simulation and curve performance evaluation described above are repeated. ``Figure 5 shows an example of a simulation in which the nuclear power plant operation automation device according to the present invention is used to increase the reactor output. In this example, an error is intentionally introduced into the core simulator. The figure shows the case where the load is added.If the output is increased according to the initial basic load curve,
Due to errors in the simulator, the measured value of the thermal limit value shows a change curve with a smaller margin than the predicted value. Therefore, when the simulator error was corrected at point B and the thermal limit value was predicted, it was found that the operation limit conditions would be violated, so an optimally corrected load curve was created by the operation plan correction section. In the optimal correction load curve, the output increase due to control rod withdrawal is interrupted at point C', and C''
By increasing the output by increasing the point reactor core flow rate, C"~C"
At point ', the output is increased again by removing the control rod. At this time, the target control rod pattern during rated operation is also modified to increase the margin of thermal limit values through predictive simulation that corrects errors.

〔Effect of the invention〕

According to the present invention, errors in predicted values of the core simulator are corrected based on actual measured values, so the prediction accuracy regarding various operating conditions when creating a load curve is improved, and there is no need to change the operating method during power operation. becomes less. Also,
When changing the driving method, the optimal road curve correction function makes it possible to change the driving method appropriately by evaluating the entire driving plan.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the basic configuration of an embodiment of the present invention;
Figure 2 is a flow diagram showing the calculation procedure of the operation plan correction section, Figure 3 is a flow diagram showing the error correction procedure of the simulator, Figure 4 is a detailed block diagram of the load curve optimum correction section, and Figure 5 is the atomic It is a chart showing an example of a simulation when the furnace output is increased. DESCRIPTION OF SYMBOLS 1... Operator, 2... Man-machine interface, 3... Plant, 4... Sub-loop control section, 5... Overall control section, 6... State prediction section, 7...
condition fIA monitoring section, 8... simulator error correction section, 9
...Curve optimum correction section, 10...Driving plan creation knowledge database. DESCRIPTION OF SYMBOLS 11... Plant data, 21... Modified load curve assumption part, 22... Input processing part, 23... Prediction simulation, 24... Result extraction part, 25...
Curve performance evaluation section, 26... Optimal measure search section, 27.
...Basic load curve, 28...Core simulator,
29... Driving restriction conditions.

Claims (1)

[Claims] 1. The overall control unit provides a control target value to the subloop control unit, and inputs control amount detection signals of each component of the nuclear power plant to the subloop control unit, so that the control amount is In a method of controlling nuclear power plant components so as to approach target values, (a) predicting the fluctuations based on the operating parameters of the nuclear power plant, and (b) determining operating limit conditions based on the above operating parameters. (c) Calculate an error correction value based on the actual measured value and predicted value of the operating parameter, (d) The determination result of (b) above and (c) ), the load curve of the operation plan database is corrected based on the error correction value of the term ), and the corrected load curve is fed back to the general control unit. 2. Calculation of the error correction value in section (c) above uses time-series data of the deviation between the actual measured value and the predicted value, creates a statistical prediction model for predicting the deviation, and uses this prediction model. The automatic operation method for a nuclear power plant according to claim 1, wherein the error correction value is calculated based on the used predicted deviation value. 3. The load curve modification in the above (d) is (i)
Using the judgment result in section (b) and the error correction value in section (c), assume a modified load curve based on the load curve in the database, and (ii) use a core simulator to calculate each of the modified load curves. (iii) the deviation between the reactor power fluctuation curve determined from the modified load curve and the required fluctuation curve, and the output fluctuation operation of the operational restriction condition margin; An optimal direction search for the modified load curve is performed using an evaluation function whose parameter is the integral value over the section, and (iv) a new modified lobe curve is assumed based on the above search result, and the optimal modified lobe curve is determined by the evaluation function. The automatic operation method for a nuclear power plant according to claim 1, characterized in that similar calculations are repeated until the card is determined. 4. A sub-loop control unit that controls each component of the nuclear power plant while receiving control amount detection signals as feedback from the component, and a control target value for the sub-loop control unit based on the load curve given from the operation plan database. (e) a state prediction unit that receives controlled variable detection signals from each component of the nuclear power plant and predicts fluctuations in the controlled variable; (f) a state monitoring unit that receives the detection signal and determines whether the controlled variable conforms to the operation restriction conditions; (g) performs a prediction based on the actual measured value and predicted value of the controlled variable; a simulator error correction unit that calculates an error correction value of the value, and (h) corrects the load curve of the operation plan database based on the determination result of the above-mentioned (f) and the error correction value of the above-mentioned (g). An automatic operation device for a nuclear power plant, characterized in that it is equipped with a curve optimization correction section.